Thursday, 1 March 2007

[CIO mind] Analyst reports: How to read them?

In this post (Analyst reports: How to read them) Felix addresses the basic rules of reading research: check who wrote it and do your homework.

We would add:

  • Understand the business model of the analyst firm (is it vendor facing or do their revenues come from end-users?)
  • Check if the report sponsored
  • Have a look at the methodology, and in particular at the sample size and distribution (did the analyst interview 5 clients in Cincinatti or did they survey 120 firms distributed in EMEA with a fair split by industry?)

    But more importantly, is the analyst riding the hype wave (like Yankee in the old days) or do they provide actionable advice?

    Special mention for market shares: whereas numbers are generally very accurate on a WW or EMEA basis, regardless of the technology. However, requesting 2 or 3 dimensional cuts from the data cube (for instance market shares for Managed Services in Denmark and Retail Banking) is asking for trouble. Do NOT EVER EVER base compensation on market share without having a discussion with the analyst on the statistical confidence interval.

    ARmadgeddon's takes:
  • The huge variability in research quality hides a constant improvement: analysts are getting better at cutting the hype and providing meaningful actionable recommendations for users. The downside is that research notes have become a bit less entertaining.
  • Don't read research naively as if it was excerpts from a holy book. Dissect, reflect and question.
  • Read also:

    No comments: